COLUMBIA GLACIER 881 fall alluded to. In 1910 the edge of the advancing glacier extended into this new lake in two lobes a short distance apart. At the western end of the upper and larger lake, which remained unchanged in 1899, 1909, and 1910, there is a higher waterfall over rock. In this upper lake the glacier margin turns from its northeast trend and continues northwestward. Here the 1892 maximum was marked by a low crescentic moraine from which the glacier had retreated so far in 1899 that there was a line of crescentic pools between the moraine and the ice front. The readvance at this point in 1909 had extended out across the site of these pools and nearly out to the 1892 terminus. We estimated the advance in this lobe to be nearly a quarter of a mile. Between August 24,1909, and June 30,1910, the continued advance of this lobe carried it out about 300 feet and there was a lateral spreading eastward of 330 feet. The 1892 moraine was completely overridden in places, so that the ice extended into the lake in three minor lobes where none reached the water ten months before. This was also true in another minor lobe to the southwest. An interesting little delta, evidently built in this lake in 1892 when the crescentic moraine was made, shows in Gilbert's 1899 photograph as a small abandoned delta with no streams except the small outlet of a marginal lakelet. In August, 1909, the ice extended nearly up to the delta, though not far enough to send streams across it; but in June, 1910, the advance had brought the ice up to it, and outwash streams were once more building it forward into the large lake. Along the more distant lobes, that project against the forested rocky knobs still farther northward, holding marginal lakelets there, the ice seemed in 1909 to extend right up to the forest and had apparently quite equalled in extent the earlier great advance in 1892. In 1910 there had been additional spreading of the glacier so that several of these lobes had advanced 200 to 400 feet into the forest. Along this lobate border there were many modifications in 1909 and 1910. The glacier had thickened, crevassing had sliced up ice that was smooth in 1899, and the distribution of moraine upon the margin had changed somewhat from 1899 to 1909. Although there was more crevassing in 1910 than in the previous year the distribution of marginal ablation moraine was about the same. The great lobe, or distributary, east of the nunatak, underwent no appreciable change from 1899 to 1910 and we assume that it still coalesces with the main glacier below the nunatak, as Palache observed in 1899. We saw this distributary from too great a distance to be sure as to the detailed conditions in 1909 and 1910, but, as the topographic map shows, it touches the main glacier but contributes no ice to it. The drainage of the eastern margin of Columbia Glacier shows an interesting alternation of complex conditions. The streams emerging from the eastern glacier distributary, joined by the drainage from the eastern margin of the main glacier south of the big nunatak, and by smaller streams from small mountain glaciers on the east, flowed southward into a lake among the rocky hills east of the main glacier. This lake was nearly filled by a delta of outwash gravels which grew forward appreciably between 1899 and 1909. From this upper lake the marginal stream flowed southward through the rocky, forested hills east of the glacier into the long, crooked, upper lake. The main mountain wall formed one shore of this lake, and the moraine bordering the minor glacier lobe that advanced a quarter mile between 1899 and 1909, formed the other. Into this muddy lake flowed numerous turbid streams from the glacier, some of which built deltas.